摘要
为克服最小方差套期保值比率参数的估计风险,本文从期货市场先验信息的设定出发,将贝叶斯统计方法应用于经典静态套期保值估计理论,系统构建贝叶斯套期保值比率估计模型,给出一类MV套期比估计新算法.选取中国商品期货市场代表性大宗商品豆粕、玉米、棉花、棕榈油;铜、铝、镍、铁矿石、螺纹钢;聚乙烯、PTA、天然橡胶,以及国际市场代表性品种COMEX白银、COMEX黄金、布伦特原油,从静态和动态两个维度开展实证分析,检验新方法的有效性.研究表明,因为利用了市场的有效先验信息,基于贝叶斯估计模型B-OLS、B-BVAR、B-ECM的套期保值效率显著优于传统的静态模型OLS、BVAR、ECM、BGARCH、ECM-GARCH;与从传统静态OLS、BVAR、ECM到动态BGARCH、ECM-GARCH两类频率统计模型的选取相比,引入先验信息所提升的套期保值效率更为显著.本文旨在从理论上提供套期保值比率研究新方法,从实践上为投资者制定最优商品期货套期保值策略提供新路径.
To overcome the estimation risk of the minimum variance hedge ratio parameter,this paper applies the Bayesian statistical method to the classical constant hedging theories based on the prior informations setting,systematically constructs the Bayesian optimal hedge ratio estimation models,and derives various mean square loss estimation formulas.Then we select soybean meal,corn,cotton,palm oil,copper,aluminum,nickel,iron ore,deformed steel,polyethylene,PTA,and natural rubber as representative commodities in China’s commodity futures market,as well as COMEX silver,COMEX gold and BCO as representative commodities in the international futures market,from static and dynamic dimensions to carry out an empirical analysis to test the effectiveness of the new models.The research shows that the hedging efficiency of Bayesian estimation models B-OLS,B-BVAR,and B-ECM based on prior informations is significantly better than the traditional five models OLS,BVAR,ECM,BGARCH,ECM-GARCH.It has also been proved that compared with the selection efficiency of traditional constant models such as OLS,BVAR,ECM and the dynamic models such as BGARCH,ECM-GARCH,the estimated efficiency improved by introducing prior information have significant advantages.This paper theoretically provides a new method to study the hedging ratio,and in practice provides a new path for investors to formulate the optimal commodity futures hedging strategies.
作者
张国胜
张怡
张顺明
陈鑫
ZHANG Guosheng;ZHANG Yi;ZHANG Shunming;CHEN Xin(School of Economics,Beijing Wuzi University,Beijing 101149,China;Software Development Center,Bank of Communications,Beijing 100031,China;School of Finance,Renmin University of China,Beijing 100872,China)
出处
《系统工程理论与实践》
EI
CSSCI
CSCD
北大核心
2023年第8期2236-2250,共15页
Systems Engineering-Theory & Practice
基金
国家自然科学基金重点项目(72233003)。
关键词
MV套期比
先验信息
贝叶斯方法
估计风险
MV hedging ratio
prior information
Bayesian method
estimation risks
